Class: WebPageArchiver::MhtmlGenerator

Inherits:
Object
  • Object
show all
Includes:
GeneratorHelpers
Defined in:
lib/web_page_archiver.rb

Overview

generates mht-files

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Methods included from GeneratorHelpers

#content_type, #download_finished?, #initialize, #join_uri, #start_download_thread

Instance Attribute Details

#conf ⇒ Object

Returns the value of attribute conf.



95
96
97
 
# File 'lib/web_page_archiver.rb', line 95

def conf
  @conf
end

Class Method Details

.generate(filename_or_uri) ⇒ String

generate mhtml (mht) file without instantiating a MhtmlGenerator object

mhtml = WebPageArchiver::MhtmlGenerator.generate(“rubygems.org/”) open(“output.mht”, “w+”){|f| f.write mhtml }

Parameters:

  • filename_or_uri (String, URI) —

    to test for

Returns:

  • (String) —

    text blob containing the result



104
105
106
107
 
# File 'lib/web_page_archiver.rb', line 104

def MhtmlGenerator.generate(filename_or_uri)
  generator = MhtmlGenerator.new
  return generator.convert(filename_or_uri)
end

Instance Method Details

#add_html_content(cid) ⇒ Object

helper method to generate proper mime part headers

param [String] cid content ID return [String] mime-part-text-blob



183
184
185
186
187
188
189
190
191
192
193
194
195
 
# File 'lib/web_page_archiver.rb', line 183

def add_html_content(cid)
  filename = File.basename(URI(@contents[cid][:uri]).path)
  @src.puts "--#{@boundary}"
  @src.puts "Content-Disposition: inline; filename=" + filename 
  @src.puts "Content-Type: #{@contents[cid][:content_type]}"
  @src.puts "Content-Location: #{@contents[cid][:uri]}"
  @src.puts "Content-Transfer-Encoding: Base64"
  @src.puts "Content-Id: #{cid}"
  @src.puts ""
  @src.puts "#{Base64.encode64(@contents[cid][:body])}"
  @src.puts ""
   return
end

#attach_contents ⇒ Object

adds mime-parts



169
170
171
172
173
174
175
176
177
 
# File 'lib/web_page_archiver.rb', line 169

def attach_contents
  #prepeare_queue
  @contents.each{|k,v| @queue.push k}
  #start download threads
  self.start_download_thread
  # wait until download finished.
  @threads.each{|t|t.join}
  @contents.each{|k,v|self.add_html_content(k)}
end

#convert(filename_or_uri) ⇒ String

convert object at uri to self-contained text-file

Parameters:

  • filename_or_uri (String, URI) —

    to test for

Returns:

  • (String) —

    text blob containing the result



113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
 
# File 'lib/web_page_archiver.rb', line 113

def convert(filename_or_uri)
    f = open(filename_or_uri)
    html = f.read
    @parser = Nokogiri::HTML html
    @src.puts "Subject: " + @parser.search("title").text()
    @src.puts "Content-Type: multipart/related; boundary=#{@boundary}"
    @src.puts "Content-Location: #{filename_or_uri}"
    @src.puts "Date: #{Time.now.to_s}"
    @src.puts "MIME-Version: 1.0"
    @src.puts ""
    @src.puts "mime mhtml content"
    @src.puts ""
    #imgs
    @parser.search('img').each{|i| 
        uri = i.attr('src');
        uri = join_uri( filename_or_uri, uri).to_s
        uid = Digest::MD5.hexdigest(uri)
        @contents[uid] = {:uri=>uri}
        i.set_attribute('src',"cid:#{uid}")
      }
    #styles
    @parser.search('link[rel=stylesheet]').each{|i|
        uri = i.attr('href');
        uri = join_uri( filename_or_uri, uri)
        uid = Digest::MD5.hexdigest(uri)
        @contents[uid] = {:uri=>uri}
        i.set_attribute('href',"cid:#{uid}")
      }
    #scripts
    @parser.search('script').map{ |i|
        next unless i.attr('src');
        uri = i.attr('src');
        uri = join_uri( filename_or_uri, uri)
        uid = Digest::MD5.hexdigest(uri)
        @contents[uid] = {:uri=>uri}
        i.set_attribute('src',"cid:#{uid}")
    }
    @src.puts "--#{@boundary}"
    @src.puts "Content-Disposition: inline; filename=default.htm"
    @src.puts "Content-Type: #{content_type(f)}"
    @src.puts "Content-Id: #{Digest::MD5.hexdigest(filename_or_uri)}"
    @src.puts "Content-Location: #{filename_or_uri}"
    @src.puts "Content-Transfer-Encoding: 8bit" if @conf[:base64_except].find("html")
    @src.puts "Content-Transfer-Encoding: Base64" unless @conf[:base64_except].find("html")
    @src.puts ""
    #@src.puts html
    @src.puts "#{html}"                      if @conf[:base64_except].find("html")
    #@src.puts "#{Base64.encode64(html)}" unless @conf[:base64_except].find("html")
    @src.puts ""
    self.attach_contents
    @src.puts "--#{@boundary}--"
    @src.rewind
    return @src.read
end